if there are factors that would tend to make early high scores overestimates, I would say that they would almost certainly be found in the "crystallized knowledge" areas of the tests, such as vocabulary, where early environmental enrichment or deprivation could make a huge difference in test performance, giving very young children with relatively enriched environments an apparent "edge" that might not be maintained once children from relatively deprived environments entered school and, hopefully, had more exposure to opportunities to acquire the same level of knowledge.
That's pretty much the way it has been explained to me as well. Vocabulary, for instance: just knowing a few "hard" words will significantly up your vocab portion of the VCI (on the WISC) at a young age b/c there is more head room. If you know more harder words b/c of an enriched environment (or fewer due to a deprived environment), that will make a more significant difference in your IQ score at a young age than it will later.
FWIW, my dd's IQ score on the WISC dropped from the 99.9th at seven to 97th at eight. I don't know that I'm sure either IQ score is correct for her b/c she has ADD and significant anxiety, but I also don't have any data that supports either score as absolutely accurate. Her WIAT scores from the time of the second WISC testing fall closer to the first IQ number but other achievement scores fluctuate wildly.
My other dd has taken multiple above level tests that pretty consistently point to her ability scores falling somewhat above the 99th percentile, but not at the 99.9th. Her IQ at 7 was just into the 99th if we add extended norms. Given all of the other data that supports that and continues to support that even as she nears her teens, I'd have no qualms about saying that her IQ was not an overestimate and was possibly a slight underestimate. My youngest I am now comfortable saying is gifted and maybe HG in some areas (she had one subtest on the WISC that was at the 99.7th and still the 99th with the second testing), but I'm not sold on saying that the highest number ever obtained is the most accurate either.
